This integrative review shows a range of implementation approaches, data collection procedures, and documentation approaches in programs for prevention of PPH at home birth using misoprostol. We recognize the limitations in comparing programs and drawing summary conclusions from different implementation models and data reporting practices, but we believe that a sufficient number of community-level misoprostol programs have been attempted to date to render discussion and interpretation of their methods and outcomes timely and appropriate. The nature and quality of the data, a majority of which was extracted from non-peer-reviewed project reports, restricts the statistical methods that could be used in data analysis, and requires the following caveats regarding generalizability.
The information that we sought to retrieve for purposes of this integrative review was not necessarily a component of the program monitoring plans for all programs, and, even if collected, was not necessarily reported or reported in a comparable manner. As a result, there are missing or assumed data for some variables of interest. For example, a common definition of PPH as an adverse event was not present in all reports, and reports that used the term excessive bleeding were assumed to be referring to perceived PPH. Explicit mention of PPH was itself absent in one report.
Additionally, this review might be biased toward more favorable results. In addition to selective data extraction from included programs, programs that were excluded from this review because of substantial missing data might have contained unfavorable results that the implementing organizations chose not to share with the public, although this is unlikely.
It is interesting to note that a substantial number of programs did not collect or report sufficient data to estimate their distribution or coverage rates. Given that misoprostol for home birth is a strategy to achieve greater protection from PPH – regardless of location of birth – we anticipated that these data would have been more readily available.
We were particularly cautious in estimating the rates of distribution and coverage of misoprostol because we understand that most programs were not attempting to reach all pregnant women within an intervention area and did not follow up with all women who received misoprostol prior to delivery. Estimations were based on available data and assumptions regarding population or sample data. The heterogeneity of program methodologies does not allow for the formation of point estimates; therefore we present rate ranges. Footnotes in the tables present additional information about calculations. Actual distribution and coverage rates at home births could be higher than those we calculated and reported.
We present misoprostol distribution separate from its coverage because fewer women might consume the drug than those who receive it. Consumption, or coverage, presents a more accurate measure of program effectiveness than distribution because it reflects both successful distribution as well as effective counseling to the woman, her family, and any involved providers.
No particular timing was predominant among programs that distributed misoprostol prior to birth (n
12), with programs using early, late, or unrestricted distribution timing. However, the range of distribution rates to the target population of pregnant women was lower for late ANC visit distribution compared to distribution at any ANC visit.
Programs that allowed distribution by CHWs and during home visits achieved greatest distribution and coverage, potentially more than double the coverage achieved by programs with distribution by health workers or as a part of ANC services. Distribution of the drug by other types of community-based workers also appeared to allow high distribution and coverage rates, in the very few programs for which this strategy is reported. This suggests that home-based distribution approaches, with relatively low-skilled providers, either singly or combined with facility-based approaches, can achieve high rates of distribution to the target population. This is potentially due to the pressures that health workers are under during their routine work and the difficulty that comes from adding additional tasks. CHWs, on the other hand, might be able to add this service to their work more easily, and likely have multiple opportunities to see a woman. As well, home-visit distribution by CHWs is primarily dependent on the actions of the worker, not the health-seeking behavior of the woman, whereas traditional ANC in a facility can only occur if the woman presents to the facility for care.
Eleven programs distributed misoprostol to women prior to birth. Several of these programs also allowed for administration to the woman at the time of birth at home, likely enhancing their overall distribution and coverage rates. The rates of ANC and skilled birth attendance are low in these program communities, so the programs strategically chose to provide women with protection against PPH even in situations where their births were not attended by SBAs.
Another area of great concern among maternal health advocates globally is whether a strategy of provision of misoprostol for home birth would detract from efforts at increasing facility birth rates. Only three of the 18 programs reviewed tracked this indicator. In none of those did the facility-based birth rate decline; indeed, the rate appeared to increase, although the calculation methods differ and the data do not conclusively support an attribution of changes to the programs themselves. Those three programs appeared to put a high value on education of the woman and her family regarding the importance of skilled attendance at birth, the dangers of PPH, and the use of misoprostol only for the situation where a woman is unable to achieve her plan of a facility-based birth.
The number of cases in which women took misoprostol prior to delivery is reassuringly low, as this is one of the areas of greatest concern for the international public health community. Administration before birth occurred in only seven cases out of more than 12,000 women who were followed up (0.06%). One case was due to a woman taking the dose before delivery of a second twin. The second twin delivered normally without complication. Another case was a woman responding to a domestic dispute with intention of self-harm. She was immediately identified and referred to a nearby facility where she delivered normally within 12 hours. Authors reporting on the Ghana program stated that there were four women who took the drug at the wrong time, three of whom took the drug after delivery of the placenta. We therefore assume that the fourth case was that of a woman who took the drug prior to birth, but no further information is available from the program description. Four cases occurred in one large program in Bangladesh for which there was no specific information about circumstances or outcomes. It is possible that there might be additional cases of administration prior to the birth that were unreported, although the likelihood of this is low, given the high profile of most of these programs.
With such a low occurrence of premature administration, it is difficult to draw any meaningful distinctions among the programs, each of which had various and unique features in design. More of the cases of premature administration occurred when the drug was distributed at any ANC visit compared to ANC or home distribution closer to the time of birth, and when distribution was by a health worker or ANC provider compared to distribution by a lay health worker.
All but one program made an attempt to identify and record the number of maternal deaths in the program’s target area, and specifically, the number of maternal deaths that occurred among women who took misoprostol. Virtually every program that recorded the number of maternal deaths also noted the method(s) by which the deaths were investigated. Investigations were also commonly undertaken to verify accounts of reports of excessive postpartum bleeding reported by women, their family, or their birth attendants. Such rigorous methods help ensure that such deaths can be more independently reviewed and evaluated for any relationship to either the drug or its method of distribution or administration. It is reassuring that there were no cases of maternal death that were attributed to misoprostol across the almost 87,000 women who took the drug as part of these programs.